Management styles and tensions in social enterprises: a quantitative approach

Social entrepreneurship has seen extraordinary growth worldwide in recent years (Defourny and Nyssens, 2010), attracting substantial attention from media, politicians, the public and academia (Pinheiro et al., 2021). Social enterprises (SEs) play a crucial role in addressing social problems with innovative solutions (Pinheiro et al., 2021). These hybrid organizations combine a social mission with financial sustainability (Nielsen et al., 2019), pursuing commercial activities to fund their goals rather than relying solely on donations or government support (Ebrahim et al., 2014). The revenues generated are reinvested to create social value for their beneficiaries (Ebrahim et al., 2014), blending the efficiency and innovation of for-profit firms with the values and mission of not-for-profits (Battilana and Lee, 2014).

Despite their potential, SEs face inherent tensions between their social and commercial missions. If poorly managed, these tensions can lead to mission drift or financial challenges (Doherty et al., 2014; Cornforth, 2014; Ebrahim et al., 2014; Nielsen et al., 2019). Additional challenges include defining success (performing tensions), managing human resources (organizing tensions), proving legitimacy to multiple stakeholders (belonging tensions) and scaling operations (learning tensions) (Battilana, 2018; Smith et al., 2013). These complexities demand a unique skill set from directors, whose management styles significantly influence decision-making and organizational performance (Ghanbaripour et al., 2020; Pereira et al., 2021; Ramos et al., 2016; Yaari et al., 2020).

Although qualitative and conceptual research on SE tensions is extensive (Pereira et al., 2021; Smith et al., 2013), quantitative studies testing how distinct management styles affect these tensions are notably limited. Understanding these effects is crucial to addressing the dual mission of SEs (Yaari et al., 2020). To address this gap, we conducted a qualitative pre-study to adapt and validate measurement scales, followed by a quantitative survey of 346 Portuguese SEs. We then test whether participative and innovative management styles are associated with lower levels of organizational tensions. This approach not only fills a quantitative void but also contributes to hybrid organizing and paradox theory by empirically assessing managerial agency in attenuating internal contradictions. SEs contribute significantly to building more equitable societies, making it essential to improve their management and sustainability.

Practically, this research may provide valuable insights for social entrepreneurs. Understanding different management styles can enhance managerial competencies and outcomes (Ramos et al., 2016). As SEs aim to support those most in need, improving their management has societal benefits. To build a more equitable world, we must help SEs thrive (Pless, 2012).

Social entrepreneurship has grown rapidly in recent years, attracting interest for its potential to tackle social problems through market-based innovation (Defourny and Nyssens, 2010; Pinheiro et al., 2021). SEs are organizations that blend a social mission with financial sustainability – using commercial activities to generate and reinvest revenues for social objectives (Nielsen et al., 2019). It is managed in an open and responsible manner and, in particular, involves employees, consumers and stakeholders affected by its commercial activities (Commission, E., 2015, p. 9). These hybrid entities pursue both social impact and financial sustainability, creating a distinctive organizational form prone to paradoxical tensions.

SEs regularly face four core tensions – performing (balancing social vs financial outcomes), organizing (structure and HR dilemmas), belonging (identity and stakeholder legitimacy) and learning (short- vs long-term horizons) – as classified by Smith et al. (2013) and refined by Civera et al. (2020). These tensions arise from the simultaneous pursuit of dual missions and pose unique challenges for SE managers. Recent comparative studies (e.g. Battilana and Lee, 2014; Hoffman et al., 2012) confirm that such tensions are central to hybrid organizing and that different managerial approaches may affect each tension type in distinct ways.

The four tension categories are summarized in Table 1, which provides a comparative overview of each tension’s description, key challenges and foundational references.

While some countries classify SEs as not-for-profit organizations, SEs are distinct in that they do not depend solely on charity or government subsidies but instead generate income through commercial activities (Battilana, 2018). Unlike traditional commercial companies, financial performance is not the ultimate goal for SEs (Alegre, 2015). Financial sustainability is necessary for SEs, but it is insufficient to fulfil their broader social mission (Alegre, 2015).

The concepts of management and leadership have been widely studied, yet their distinctions remain debated. Kotter (1990, 1995) highlights that management is tactical, producing order and consistency, while leadership is strategic, enabling change and innovation. Maccoby (2000) differentiates management as a functional role in business and leadership as a relational dynamic that energizes organizations. However, Nienaber (2010) suggests that management and leadership tasks are intertwined, with leadership overlapping management. Ultimately, both are complementary and essential for organizational success (Algahtani, 2014).

Our conceptualization of management styles in SEs draws on three complementary models, all rooted in hybrid organizing and paradox theory (Smith and Lewis, 2011; Besharov and Smith, 2014). Firstly, Adizes, (1976) classical roles – productive (focus on outcomes), administrator (structure and control), entrepreneur (innovation and change) and integrator (people and cohesion) – provide a rich typology for profiling managers. Secondly, Klijn et al. (2008) emphasize the importance of stakeholder interaction, strategic orientation and process flexibility – key in hybrid settings where leaders must continuously balance internal and external demands. Thirdly, Olmedo–Cifuentes and Martínez-León’s (2014) simpler continuum (participative vs competitive) translates these roles into two operational styles that map directly onto SEs’ tensions: inclusive, consensus-oriented decision-making versus autocratic, task-driven approaches.

Empirical studies have applied these models across contexts – from work-integration SEs in Israel (Yaari et al., 2020) to project management in Brazil (Ramos et al., 2016), Oman (Chaudhry et al., 2019) and construction (Ghanbaripour et al., 2020) – demonstrating their generalizability. In SEs, where the dual mission creates performing, organizing, belonging and learning tensions (Smith et al., 2013; Civera et al., 2020), participative styles (democratic, relationship-focused; Dutot, 2017) are expected to mitigate belonging and organizing tensions, while entrepreneurial or “innovative” styles (proactive, creativity-oriented) should alleviate performing and learning tensions by enabling adaptive processes and novel solutions.

Recent international studies enrich the understanding of leadership in SEs and reinforce the relevance of participative and values-driven leadership in SEs. Studies by Chang and Jeong (2021) and Brunelli and Cavazotte (2024) highlight how context-sensitive and relational leadership approaches contribute to improved performance and help address the inherent tensions of hybrid organizations.

Beyond foundational accounts, recent work sharpens how leadership interacts with hybrid governance and institutional logics. Reviews and syntheses document the field’s evolution and boundaries of leadership effects within hybridity, including hybrid organization stock-takes, hybrid board governance challenges and hybrid value logics that structure decision tensions. Emerging studies also highlight cross-sector collaboration as a governance lever to manage tensions. Together, these advances suggest that while leadership matters, it likely accounts for a small share of variance amidst structural and contextual drivers (Banerjee et al., 2024; Ratinho and Bruneel, 2024 and Jakob and Sundermeier, 2025).

Building on organizational tensions and management‐style models, we ground our framework in paradox theory (Smith and Lewis, 2011; Besharov and Smith, 2014). Paradox theory underscores that hybrid organizations must navigate interdependent and persistent tensions – performing, organizing, belonging and learning – that arise from the simultaneous pursuit of social and economic goals. Managers exercise agency through distinct styles, which can modulate how these tensions are experienced and managed.

We integrate three complementary lenses on managerial agency: Adizes, (1976) roles offer archetypes for how managers allocate focus between outcomes, structure, innovation and people, Klijn et al.’s (2008) stakeholder‐interaction model emphasizes flexibility and continuous negotiation with internal and external actors, crucial for hybrid logics; and Olmedo-Cifuentes and Martínez-León’s (2014) continuum distils these profiles into two operational styles that align with the four tension types.

We theorize two leadership–tension pathways grounded in paradox and hybrid organizing. Firstly, participative styles reduce belonging and organizing tensions by strengthening voice and identification, clarifying roles and processes and enabling inclusive governance. Secondly, innovative/entrepreneurial styles attenuate performing and learning tensions by fostering experimentation, adaptive routines and ambidexterity that balance short- and long-term objectives (Pearce et al., 2019). These mechanisms convert descriptive links into causal logics to be tested quantitatively in SEs.

Figure 1 presents a higher-order structural equation modelling (SEM). Here, “Management Styles” and “Tensions in SEs” are treated as second-order constructs, measured by first-order constructs and variables (not shown in Figure 1 for clarity). This framework allows us to test not only whether management styles matter (H1) but also – via second-order SEM – the relative strength of each style’s moderating effect on SE tensions.

We propose the following hypothesis:

We can formulate these sub-hypotheses:

We adopted a sequential exploratory–quantitative design. Firstly, an exploratory qualitative pre-study was conducted to adapt and validate our measurement scales in the SE context. Secondly, a large-N survey was administered to 346 Portuguese SEs, with data analysed using PLS-SEM. This approach ensures content validity of the instrument and leverages quantitative rigor to test our hypotheses.

Prior to the survey, we conducted four semi-structured interviews (45–60 min each) with directors of Portuguese SEs. The interview protocol comprised three sections: organizational context, with questions about the organization’s mission, business model and the interviewee’s role; management style, focusing on how the interviewee manages people, sets goals and makes decisions; and types of tensions experienced within the SE. All interviews were audio-recorded and coded using a template-thematic approach (Pattinson, 2025). The insights were used to adapt wording, eliminate redundant items and add contextually relevant probes. Given the exploratory purpose (rather than theory building), thematic saturation was reached quickly, which we acknowledge as a limitation. Beyond the reliability and validity checks reported in the PLS-SEM section (e.g. CR, AVE, HTMT), we complemented interviews with expert review by two SE scholars to assess content adequacy. We make no claims of thematic saturation.

Insights from the interviews were used to develop the questionnaire, which was also structured in three parts. The first part captured the social entrepreneur’s profile and their management style, exploring their approach and role within the organization. The second focused on identifying the tensions encountered in the organization’s daily operations. The third collected demographic details and information about the organization’s profile, along with personal details of the social entrepreneur. A detailed description of the questionnaire structure is presented in Table 2.

The statistical approach used to evaluate the sample was partial least squares SEM (PLS-SEM). This method was chosen because it is particularly suitable when the theoretical foundation is less established and the data are not normally distributed. Compared to other approaches such as regression and covariance-based SEM (CB-SEM), PLS-SEM is more robust and provides more reliable estimates of population parameters under these conditions (Hair et al., 2021).

The Henze–Zirkler multivariate normality test confirmed that the items in this study do not follow a normal distribution, which is a requirement for CB-SEM. Additionally, PLS-SEM is well-suited for estimating complex cause–effect relationships in diverse research contexts. Its flexibility, including fewer restrictions on the number of variables used to measure a construct, makes it ideal for this study (Hair et al., 2021).

To assess the management style of each manager, an instrument adapted from the one proposed by Ramos et al. (2016) was used. This instrument includes statements about how managers act in their daily activities, considering various dimensions. The questionnaire’s second section (Section 2) has been validated by previous studies and used by other authors, such as Chaudhry et al. (2019) and Ghanbaripour et al. (2020). While the original version consisted of 29 questions, it was adapted to the context of SEs, resulting in a revised version with 27 questions, as shown in Table 2.

Each item required respondents to indicate the frequency of their behaviors using a five-point Likert scale, where 1 represented “never” and 5 represented “very regularly, if not always”. This approach was designed to measure the intensity of each dimension within the social manager’s profile. Table 3 provides a detailed description of all the questions and variables.

As previously discussed, SEs are complex organizations that encounter various tensions throughout their lifecycle. To evaluate the intensity of the four types of tensions identified in the literature (Smith et al., 2013), the framework proposed by Civera et al. (2020) was used to design Section 3 of the questionnaire. This section comprises 15 questions.

Managers were asked to assess the intensity of each tension within their organization using a five-point Likert scale, where 1 represented “never” and 5 represented “very regularly, if not always”. Table 4 provides a detailed overview of the dimensions associated with each type of tension.

The sample for this study was constructed using data from the Base de Dados Social and the Mapa Social. Despite the significant role of the social sector in Portugal’s economy, existing data sources often provide fragmented and limited information with low levels of interaction and poor visualizations. To address this gap, Nova SBE Data Science Knowledge Centre, in partnership with the “laCaixa” Foundation, BPI and Nova SBE, launched the Base de Dados Social in November 2020. This online platform aims to provide integrated and dynamic access to information about Portuguese social impact organizations.

From the Base de Dados Social, 16,299 results were exported, but only 3,736 included an email contact. To strengthen the database, additional data was extracted from the Mapa Social, which consolidates information about institutions and their social responses in Portugal. This source yielded 11,583 results, of which 6,156 included email addresses. After removing duplicate contacts, the final database comprised 8,750 general email addresses.

Despite the large number of results, not all organizations were suitable for the study. Since the legal status of “social enterprise” is not yet consolidated in Portugal, accurately targeting SEs in the initial phase was challenging. To address this, the first question of the questionnaire asked respondents whether their organization aligned with the European Commission’s definition of SEs, as outlined in the Portugal country report on Social Enterprises and their Ecosystems in Europe (Commission E., 2015). Organizations that did not meet this criterion were excluded from the study.

Following this screening, a total of 1,870 emails were sent, containing a link to the questionnaire via Google Forms. Of these, 386 responses were received, and 346 were deemed valid for the study, resulting in a response rate of 18.5%. However, as some email addresses were inactive or outdated, the actual response rate is likely underestimated.

To evaluate the hypotheses of this study, a PLS-SEM was implemented in R using the SEMinR 2.3.1 package. As outlined in the theoretical framework, the constructs Management styles and tensions were treated as second-order constructs due to the high correlations observed between both the second- and first-order constructs. Table 5 presents the Spearman’s correlation coefficients for management styles and tensions. The significant correlations in the table indicate that considering these constructs as second-order variables is reasonable.

As noted by Hair et al. (2021), using higher-order structural models for complex constructs enhances parsimony and improves interpretability. The framework of this study is modelled based on a reflective measurement model, which assumes that the construct drives the measurement of its indicator variables (Hair et al., 2021). Consistent with theoretical expectations for reflective models, causation flows from the second-order latent variable to the first-order latent variables and subsequently from the first-order latent variables to their respective items.

The structural model estimates were evaluated only after confirming the reliability and validity of the constructs. Consistency reliability assesses how consistently a set of items measures a construct. While Cronbach’s alpha (CA) is commonly used, it tends to underestimate internal consistency reliability. Composite reliability (CR), ranging from 0 to 1, is a more suitable metric, with values above 0.6 acceptable for exploratory research, 0.7–0.9 considered good, and values over 0.95 problematic (Hair et al., 2021).

Convergent validity evaluates the extent to which measures of the same construct correlate. This is assessed via the outer loadings of indicators and the average variance extracted (AVE), which should exceed 0.5 to indicate that the construct explains at least half the variance of its items (Hair et al., 2021).

Discriminant validity measures whether a construct is empirically distinct from others. While the Fornell–Larcker criterion compares a construct’s AVE to its squared inter-construct correlations (Fornell and Larcker, 1981), Henseler et al. (2015) demonstrated its limitations and proposed the heterotrait-monotrait ratio (HTMT) instead. Discriminant validity issues exist if HTMT values exceed 0.90 for conceptually similar constructs or 0.85 for more distinct constructs. Additionally, the HTMT confidence interval (based on bootstrapping with 10,000 subsamples) must not include 1 (Henseler et al., 2015; Hair et al., 2021).

Low-loading items (below 0.40) were eliminated to improve reliability and validity, following Hair et al. (2021). Constructs “Competitive Style” and “Conservative Style” were removed due to insufficient items after elimination. Items “PTS6” and “PTS7” from “Participative Style” and “INVS2”, “INVS4”, “INVS5” and “INVS6” from “Innovative Style” were excluded for similar reasons. Items with loadings between 0.40 and 0.70, such as “INVS6” and “PTS5”, were retained because they did not improve CR or AVE upon removal and were close to the acceptable threshold of 0.70. In social sciences, such weaker loadings are common (Hulland, 1999).

Following these adjustments, Table 6 confirms the model’s reliability and validity. CR values for all constructs met the required thresholds, and AVE values demonstrated that constructs explained more than 50% of the variance in their indicators, satisfying criteria from Bagozzi and Yi, (1988) and Fornell and Larcker (1981).

Discriminant validity was assessed using the HTMT ratio and bootstrapped confidence intervals, as shown in Table 7. None of the HTMT values exceeded 0.85, and the confidence intervals did not include 1, indicating that the constructs are distinct (Henseler et al., 2015; Ullah et al., 2020).

In conclusion, the model meets the required standards for validity and reliability.

After removing constructs and items that did not meet validity and reliability criteria, the structural model was re-estimated using the remaining constructs and items. In this final stage, the hypothesized relationship between management styles and tensions (H1) was evaluated. The structural model was assessed for explanatory power by calculating the coefficient of determination (R²) for endogenous constructs (see Figure 2). The R² for the structural path was 0.058, indicating low explanatory power. However, R² should always be interpreted in the context of the study, considering that it is influenced by the number of predictor constructs; a greater number of predictors typically results in higher R² values (Hair et al., 2021).

Next, the model and hypotheses were tested using bootstrapping with 10,000 subsamples. The results of the analysis are presented in Figure 2. The relationship between management styles and tensions is summarized in Table 8. The estimated coefficient indicates a negative relationship, demonstrating that higher adoption of certain management styles is associated with lower levels of tensions in SEs.

The findings provide preliminary support for the hypotheses (H1 and H1a, H1b and H1c). The relationship between management styles and tensions is both statistically significant and negative, as indicated by the confidence interval, which includes only negative values. This result suggests that management style has a negative impact on tensions in SEs. Specifically, adopting a more participative (H1b) and innovative management style (H1c) is associated with reduced levels of tensions within these organizations. The evidence is preliminary and non-causal.

As post-hoc, exploratory probes, we examined each style’s relationship with the four tension dimensions. Participative management showed its strongest negative associations with belonging – (β = −0.30) and organizing tensions (β = −0.28), indicating that inclusive decision-making and stakeholder engagement particularly ease identity and structural conflicts. Conversely, innovative (entrepreneurial) style had more pronounced effects on performing (β = −0.27) and learning tensions (β = −0.25), suggesting that creativity and adaptability are most effective at mitigating goal-setting and growth-related paradoxes. These differentiated patterns highlight that training programmes might focus participative practices on identity/structure issues, while innovation workshops could target performance/learning challenges.

These findings align with paradox theory (Smith and Lewis, 2011; Besharov and Smith, 2014), which argues that hybrid organizations face multiple, interdependent tensions. That participative styles best reduce belonging and organizing tensions suggests that collaborative, inclusive leadership directly addresses identity‐ and structure‐related paradoxes, while innovative styles mitigate performing and learning tensions by introducing adaptive and creative problem-solving, further validating our integrated framework.

This study focused on SEs, analysing how their management styles influence the tensions they face. Taken as exploratory evidence, the relationship is statistically significant yet modest (R² = 0.058)

The descriptive analysis of the sample reveals noteworthy trends. Nearly 60% of the population in our sample are female. According to Pordata data from 2021, women constitute 49% of Portugal’s active population and hold only 27% of managerial positions. This contrast highlights an interesting disparity, as women appear to dominate leadership roles in SEs. Further research could explore why women hold more leadership roles in SEs and whether this pattern is consistent across contexts.

The descriptive analysis also shows that the innovative management style is the one most commonly identified by managers in our sample. This is consistent with past research highlighting creativity and innovation as key to managing challenges and tensions in SEs (Gidron, 2014; Kerlin, 2006; Yaari et al., 2020). Regarding tensions, the performing tensions presented the highest mean, consistent with earlier studies (Battilana and Lee, 2014; Nielsen et al., 2019; Smith et al., 2013). This underscores the prominence of challenges such as evaluating overall performance and avoiding mission drift in the daily realities of SEs.

Our adaptations of Ramos et al. (2016) to assess management styles proved useful, though they could benefit from further refinement. Reducing the constructs to two for validity reasons suggests that the original questionnaire may not fully reflect the complexity of SEs. On the other hand, the questionnaire designed to assess tensions in SEs, primarily based on Smith et al. (2013), met all validity requirements, demonstrating its suitability for this context.

The results of our hypothesis testing are consistent with prior research emphasizing the pivotal role of social entrepreneurs in mitigating tensions within SEs (Cornforth, 2014; Yaari et al., 2020; Smith et al., 2012; Tracey and Phillips, 2007). The key finding is that participative and innovative management styles help reduce tensions in SEs.

This study is innovative because, to our knowledge, it is the first to quantitatively combine the concepts of SEs, management styles and tensions. These results deepen our understanding of SE dynamics and offer actionable guidance to improve their management practices.

It is also important to emphasize that the model’s explanatory power, as measured by the R² value (0.058), is relatively low. While this is not uncommon in social science research (Ozili, 2023), particularly in exploratory studies with complex constructs, it indicates that management styles alone do not fully account for the variance in organizational tensions. Theoretically, this suggests that leadership agency is only one of multiple drivers of tensions in SEs; structurally embedded factors – governance design, stakeholder complexity, external environment – likely account for the remaining variance. Practically, it cautions against over-reliance on leadership training alone and points to the need for integrated interventions across strategy, governance and measurement systems. Future research could consider additional explanatory variables, such as organizational culture, leadership experience, stakeholder complexity or external environmental pressures. Incorporating these dimensions may help build more comprehensive models and deepen our understanding of the multifaceted sources of tension in SEs.

In this study, we placed SEs at the centre of analysis to understand how managerial agency shapes the paradoxes inherent in their hybrid missions.

Despite these contributions, our cross-sectional design and the modest R² (0.058) indicate that while leadership matters, structural and contextual factors likely drive the bulk of tension variance. While prior work highlights leadership agency in hybrids, our results suggest leadership likely accounts for a small portion of tension variance, consistent with views that structural and environmental factors dominate; thus, we position leadership as necessary but not sufficient.

Our descriptive analysis – showing a majority of female managers (60% vs 49% nationally) and a predominance of innovative leadership styles – underscores both the distinct profile of SE leadership and the centrality of creativity in tension management (Gidron, 2014; Kerlin, 2006). Future work should unpack the institutional and cultural drivers of gender representation and style adoption across varied contexts.

Our adapted measure of management styles proved reliable, although the need to collapse constructs suggests further refinement is needed to capture SE complexity. By contrast, our tension scale demonstrated robust validity, providing confidence in our multidimensional findings.

This research extends paradox theory into the SE domain and offers starting points for assessing leadership effectiveness. Nonetheless, the limited scope of our qualitative pre-study (n = 4) and the Portuguese‐only sample constrain generalizability. Future research should adopt longitudinal, multi-country designs to test the robustness of these relationships. We also recommend that future research use a broader qualitative base (e.g. 15–30 interviews across organizational archetypes) prior to survey validation.

While SEs are often associated with philanthropy, financial sustainability is essential for their survival and long-term social impact. Without adequate funding, their ability to create meaningful change will diminish. As a society, we must recognize the importance of SEs and encourage governments, the private sector and future generations to support their growth. We hope this study inspires action because, as the saying goes, love does not pay the bills. SEs are vital to our society today and even more so for tomorrow. It is up to all of us to ensure their success.

The findings of this study carry important implications for both public policy and the broader field of social innovation. Preliminary patterns suggest that participative and practices may be most relevant where belonging/organizing tensions prevail, and innovative practices where performing/learning tensions dominate; capacity-building can therefore be structured as targeted experiments with clear outcome metrics. Public policies aimed at supporting the professionalization of leadership in SEs – such as targeted training, mentorship schemes or incentives for collaborative governance – can enhance organizational resilience and reduce mission drift. Given the modest association found in this exploratory study, managerial and policy implications should be treated as hypothesis-generating and subject to evaluation.

Moreover, funding mechanisms and social investment tools could prioritize or condition support to SEs that adopt inclusive and innovative governance models, thereby aligning financial assistance with improved organizational practices. Pilot programmes with monitoring to assess whether participative modules measurably reduce belonging/organizing tensions, and innovation modules affect performing/learning tensions.

Future research can adopt a multi‐method approach, beginning with longitudinal studies over three to five years to assess how targeted leadership training influences tension reduction and mitigates mission drift over time. In another direction, comparative analyses that contrast SEs with for-profit SMEs on talent attraction and retention strategies could help isolate the sector-specific factors driving leadership dynamics. Cross-cultural replications in diverse national contexts are needed to determine whether the predominance of female leadership observed in Portuguese SEs generalizes to other settings.

The data that support the findings of this study are available from the corresponding author upon request.

This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sectors.

During the preparation of this work, the author(s) used ChatGPT4.0 to correct the English. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.